Theoretical Lower Limit of Coercive Field in Hafnia

Hafnia-based ferroelectrics hold great promise for next-generation electronic devices, but their high coercive field (E_C) remains a key challenge. In this talk, I will present a deep-learning-assisted multiscale framework that integrates atomistic insights to uncover the mechanisms behind ferroelectric switching in hafnia. Focusing on both nucleation-limited switching (NLS) and domain-wall-driven Kolmogorov-Avrami-Ishibashi (KAI) models, our approach accurately predicts the corresponding coercive fields. Our NLS-type E_C predictions match well with experimental data, including our own measurements, and correctly capture the observed thickness dependence in ultrathin films (3-20 nm). We show that the giant E_C values in hafnia arise from the ultrathin geometry, which restricts switching to the NLS regime. Furthermore, we identify a theoretical lower bound of 0.1 MV/cm for KAI-type switching, enabled by mobile domain walls. Together, these results offer a unified understanding of switching phenomena in hafnia and pave the way for low-E_C ferroelectric devices.